Predictive Analytics & Machine Learning Foundations
Build predictive-analysis judgment: use proper splits, baselines, and evaluation to decide whether a model is useful rather than merely complex, from target, baselines, train/validation/test and leakage, through regression and classification metrics, cross-validation, feature importance and fairness, to error analysis, drift and deployment decisions, then complete your own Predictive Analytics Use Case + Model Card.
The course storyline: from framing a prediction problem to a deployment decision
The whole course follows one line: Frame, Baseline, Split, Train, Evaluate, Error Analysis, Deployment Decision. AI may help write code; people stay responsible for verification and the final judgment.
- 1. Frame
What is the prediction unit, target, and decision?
- 2. Baseline
How well can you do without machine learning?
- 3. Split
Are train/validation/test split correctly, with no leakage?
- 4. Train
Is the model trained at a reasonable complexity, without overfitting?
- 5. Evaluate
Does the model clearly beat the baseline? By which metric?
- 6. Error Analysis
Does error concentrate in a specific subgroup or situation?
- 7. Deployment Decision
Is this result worth deploying? Are cost, risk, and monitoring in place?
Your next step
DS 201 · Module 1: Frame a Prediction Problem and Build a First Model
Curriculum
- Module 1Frame a Prediction Problem and Build a First ModelStart with target, features, and prediction time, then learn baselines, train/validation/test, and leakage, and build your first verifiable predictive model in a Guided Lab.
- Module 2Evaluate, Interpret, and Decide on DeploymentLearn classification and thresholds, cross-validation and tuning, feature importance and fairness, then error analysis, drift, and deployment decisions, and complete an Independent Lab Use Case.
Progress is stored in this browser and does not identify you. It is uploaded only if you join progress sync below.
Sync my progressJoin a session to keep progress across devices and share it with your instructor
Practice and assessment
- Practice A: Prediction Framing CanvasDefine the prediction unit, target, decision, and columns available at prediction time.Not done yet
- Practice B: Leakage HuntFind the column in the data that is target leakage disguised as a feature.Not done yet
- Practice C: Baseline ChallengeBuild a baseline for both the regression and classification targets, with reasoning.Not done yet
- Practice D: Metric SelectionJudge whether a business scenario should prioritize precision or recall.Not done yet
- Practice E: Threshold Trade-offCompare the false-alarm vs. missed-case trade-off at different thresholds.Not done yet
- Practice F: Error Slice ReviewCheck whether model error concentrates by subgroup, and raise questions to investigate before deployment.Not done yet
- DS 201 pre-test and post-testPre- and post-tests check how your judgment changes on target, baselines, splits and leakage, regression and classification metrics, thresholds and CV, importance, and deployment decisions.Not done yet
- Predictive Analytics Use Case + Model CardTurn what you learned into your own Predictive Analytics Use Case + Model Card, stating the problem, data, method, verification, and deployment decision.Not done yet
About this course
Outcomes
- Define prediction unit, target, horizon, decision, and information available at prediction time
- Use naive/majority/mean or rule-based baselines before ML
- Separate fitting/tuning/final evaluation; prevent future/target leakage
- Fit a simple regression model; interpret MAE/RMSE in domain units
- Build baseline and model with a fixed split; compare held-out performance
- Complete the final artifact: Predictive Analytics Use Case + Model Card
Prerequisites
- DS 104 recommended
- DS 101–102 level Python/pandas skills expected
Who it is for
- NTPU students, faculty and administrative staff who work with structured data
- Learners who completed DS 104 and want to build predictive-analysis and machine-learning judgment
- Analysts who need stronger verification, documentation, and responsible AI/data practice
Certificate
Earns a course completion certificate (L1); a capability certificate (L2) is also available through additional artifact review.
Requirements to earn it
- Complete all 2 course modules
- Pass the post-test assessment at 80% or higher
- Submit a project artifact and pass review
Course feedback
It takes 2 to 3 minutes and is used to improve the course.